TheEuropeanCommissionand the Integration of Europe

The European Commission andthe Integration of EuropeImages of Governance

Liesbet Hooghe

PUBLISHED BY THE PRESS SYNDICATE OF THE UNIVERSITY OF CAMBRIDGE

The Pitt Building, Trumpington Street, Cambridge, United Kingdom

CAMBRIDGE UNIVERSITY PRESS

The Edinburgh Building, Cambridge CB2 2RU, UK40 West 20th Street, New York, NY 10011-4211, USA477 Williamstown Road, Port Melbourne, VIC 3207, AustraliaRuiz de Alarcon 13, 28014 Madrid, SpainDock House, The Waterfront, Cape Town 8001, South Africa

http://www.cambridge.org

C© Liesbet Hooghe 2001

This book is in copyright. Subject to statutory exceptionand to the provisions of relevant collective licensing agreements,no reproduction of any part may take place withoutthe written permission of Cambridge University Press.

First published 2001

Printed in the United Kingdom at the University Press, Cambridge

Typeface Plantin 10/12 pt. System LATEX 2ε [TB]

A catalogue record for this book is available from the British Library.

ISBN 0 521 80667 4 hardbackISBN 0 521 00143 9 paperback

Contents

List of figures page viList of tables viiPreface ix

Prologue 1

1 Preference formation in the European Commission 6

2 Men (and women) at Europe’s helm 31

3 Images of Europe 67

4 Beyond supranational interest 93

5 Capitalism against capitalism 118

6 Principal or agent 142

7 Accommodating national diversity 168

8 Conclusion 193

Appendix I: Statistics 216

Appendix II: Description of independent variables 224

Appendix III: Survey material 233

References 249Index 273

v

Figures

1.1 Top officials and preference formation onEU governance page 25

4.1 Hypotheses on top officials’ preferencesfor supranationalism 100

5.1 Hypotheses on top officials’ preferences for Europeanregulated capitalism 127

6.1 Top officials and types of international governance 1486.2 Hypotheses on top officials’ preferences for

administrative management 1567.1 Hypotheses on top officials’ preferences for consociational

accommodation 1808.1 Top officials and the Commission’s institutional interest 1968.2 Top officials and preference formation on

EU governance 210

vi

Tables

2.1 Officials by DG grouping page 432.2 Officials by nationality grouping 442.3 Interest in European politics 522.4 Men and women interviewees 532.5 Study abroad for top officials 542.6 DG experience by seniority 562.7 Top officials’ prior work experience 572.8 Recruitment into the Commission 602.9 Party identification and political activity 612.10 Time budget of a typical top Commission official 622.11 Frequency of contact with actors inside the Commission 632.12 Frequency of contact with EU institutional actors and

governmental and societal actors 642.13 Frequency of contact between societal groups by DG

grouping 653.1 Indicators for factor analysis 683.2 Factor analysis of attitude indicators for Commission

officials 733.3 Images of Europe – where Commission officials stand 763.4 Role definitions of top officials 864.1 Top officials on supranationalism 1074.2 Explaining supranationalism: multivariate OLS regression 1094.3 Supranationalism by party identification 1114.4 Four variables that capture variation among nationalities 1145.1 Top officials on European regulated capitalism 1235.2 Explaining European regulated capitalism: multivariate

OLS regression 1345.3 European regulated capitalism by party identification 1355.4 European regulated capitalism by type of capitalism

and corporatism 1376.1 Correlations: dimensions of international governance 1486.2 Top officials on administrative management 161

vii

viii List of Tables

6.3 Explaining administrative management: multivariateOLS regression 162

6.4 Delors recruits on administrative management acrosssubgroups 164

7.1 Top officials on nationality in the Commission 1797.2 Explaining consociational accommodation: multivariate

OLS regression 1857.3 Consociational accommodation for internal recruits and

parachutists by cabinet experience 1868.1 Explanatory power of external and internal models of

preference formation 2138.2 Explanatory power of socialization and utility models of

preference formation 214A.1 Officials by DG in the Commission 216A.2 Officials by nationality 217A.3 Personal and professional characteristics of top officials 218A.4 Dependent variables: descriptive statistics 219A.5 Dependent variables: correlations 219A.6 Independent socialization variables: descriptive statistics 220A.7 Independent socialization variables: correlations 221A.8 Independent utility maximization variables: descriptive

statistics 222A.9 Independent utility maximization variables: correlations 223

1 Preference formation in the EuropeanCommission

Senior civil servants work in two worlds simultaneously: they make andtake orders for routine actions in the hierarchical world of public admin-istration, on the basis of formal rules, cost–benefit analysis, and exper-tise; and they mobilize support for contentious decisions in the non-hierarchical world of politics, through networking and arm-twisting.1

Combining politics with expertise is indispensable in modern governance.Yet the relationship between these two worlds is delicate and contested,and senior civil servants, who are the interface between them, are directlyaffected.

Several institutional characteristics of the European Union exacerbatethe tension between political agenda-setting and bureaucratic service(Hooghe 1997; Page 1997). In conjunction with the College of Commis-sioners, the officials of the European Commission have a constitutionalobligation to set the legislative agenda because they have exclusive formalcompetence to draft EU legislation (except in foreign policy and some asy-lum and immigration issues). This competence sets senior Commissionofficials apart from their counterparts in national administrations. So theypromote the policies of their directorate to private interests, politicians,public, and, last but not least, reluctant Commission colleagues. Theydirect negotiations between the Commission, on the one hand, and theCouncil of Ministers’ working groups, the European Parliament, and in-terest groups, on the other. They broker legislative negotiations betweenthe Council of Ministers and the European Parliament. Yet, as careercivil servants, they also execute and administer political decisions takenby elected leaders. In that capacity, they provide administrative and man-agerial leadership to over 4,000 Commission administrators. Second, themutual responsibilities of national governments and supranational insti-tutions over legislation are contested. The European Union does not have

1 Several researchers have studied hybrid bureaucratic and political roles for national ad-ministrations in advanced industrial societies (Aberbach, Rockman, and Putnam 1981;Campbell and Peters 1988; Christensen 1991; Ingraham 1998; Page 1985; Suleiman1975; Suleiman and Mendras 1995; Wood and Waterman 1993).

6

Preference formation in the European Commission 7

a single political executive. The notion that the European Commissionis merely an agent of national governments is complicated – and weak-ened – by competition among the European Parliament, the EuropeanCouncil, and the various functional Councils of Ministers. Interlockingcompetencies create significant scope for autonomy for senior Commis-sion officials, though it also exposes them to the criticism that they are“a run-away bureaucracy” (Pollack 1997).

Commission officials are just one example of an expanding group of un-elected appointees empowered to take authoritative decisions. Like pub-lic regulators, central bankers, intergovernmental negotiators, supremecourt judges, and officials in international organizations such as the WorldTrade Organization, the World Bank, or the International MonetaryFund, Commission officials combine limited democratic accountabilitywith authority that rests on expertise, partisan impartiality, and dele-gated competence. Unelected political actors walk a fine line between“being political” – which means making value-based choices – and “be-ing an expert” – which is presumed to lift them above partisan choice.They undermine their legitimacy if they confound these roles. But this issometimes difficult to avoid.2

In the complex setting of the European Union, Commission officialsoften find it impossible to resolve the tension between politics, expertise,and impartiality. As “guardians of the Treaties” endowed with uniquepowers of legislative initiative and considerable executive powers over anextraordinarily broad range of issues, they are intimately involved in mak-ing authoritative, that is, political, decisions (Noel 1973). As a body ofunelected officials appointed for their expertise, the authoritative powerof the Commission is second to none in contemporary advanced democ-racies. Yet, to the extent that their decisions are perceived as breachingimpartiality, they are open to criticisms of partiality.

2 A dramatic example is the US Supreme Court’s final ruling in the presidential contestbetween George W. Bush and Al Gore of December 12, 2000, which split the Court alongideological lines. Leading commentators – left and right – decried the politicization ofthe Supreme Court. The Justices tread the political terrain with some trepidation. Theyprefaced their ruling with a reminder to the public of the delicacy of their task: “Noneare more conscious of the vital limits on judicial authority than are the members of thisCourt, and none stand more in admiration of the Constitution’s design to leave theselection of the President to the people, through their legislatures, and to the politicalsphere. When contending parties invoke the process of the courts, however, it becomesour unsought responsibility to resolve the federal and constitutional issues the judicialsystem has been forced to confront” (George W. Bush et al. vs. Albert Gore, Jr., et al., 531U.S. 949 (Dec. 12, 2000).) In his dissenting opinion, however, Justice John Paul Stevensbluntly conceded that the courts had failed to walk the fine line: “Although we may neverknow with complete certainty the identity of the winner of this year’s presidential election,the identity of the loser is perfectly clear. It is the nation’s confidence in the judge as animpartial guardian of the rule of law.”

8 The Commission and the Integration of Europe

Senior Commission officials have always been conscious of their vul-nerable position at the intersection of politics and expertise. But theirposition has become more precarious as European integration has deep-ened. In June 1992, a sliver-thin majority of Danish citizens rejected theMaastricht Treaty, which laid the basis for a currency union by 1999and devolved more national competencies to the European institutions.This sent a shockwave through the European political class. The initialrejection was reversed in the second Danish referendum in 1993, but therules of the game had changed. The permissive consensus in favor ofdeepening European integration was shattered, and replaced by a con-straining dissensus. The extent to which national sovereignty should bediluted and market integration should be complemented with European-wide political regulation is now a matter of dispute among governments,political parties, and citizens. The action has shifted from near-exclusiveinteractions between national governments and technocrats to encompasspolitics in the usual sense: party programs, electoral competition, parlia-mentary debates and votes, public opinion polls, and public referenda.Elitist decision-making has come to an end.

The Commission has taken the brunt of the blame for public uneasewith European integration. It reached its absolute low point in publicesteem in Spring 1999, when the College of Commissioners resigned inthe face of allegations of nepotism, fraud, and mismanagement of funds(see chapter 6). Subsequent investigations into management practices inthe bureaucracy provided the context for a top–down reorganization ofCommission services and a reshuffling of top officials. These brisk mea-sures shook the confidence of many Commission officials to the core, andthey did little to clarify the Commission’s uneasy balancing act betweenpolitics, expertise, and impartiality.

Top Commission officials, then, are by no means above or beyond thefray of EU politics. Given the powers and responsibilities they have, theyare drawn into debates on the chief issues facing the Euro-polity. Andtheir role as the interface between politics and bureaucracy is directlyaffected by the transition in European governance from a largely func-tionalist, technocratic system for interstate collaboration to a Europeanpolity where objectives and decision rules are openly contested. Where dothese people stand on the issues that shape the organization of authorityin the European Union? What are their political preferences on Europeangovernance?

Questioning top Commission officials

In the European Union, a political struggle is being waged about first prin-ciples of political authority. It revolves around four enduring questions,


which structure this book. How should authority be structured acrossterritorial layers of government? What should be the scope of public au-thority in the economy? Should the European elite be democraticallyaccountable? What role(s) should organized interests play in governingEurope? Let us examine these in turn.

What should be the primary locus of authority?

European integration has been undertaken on pragmatic grounds in pur-suit of collective benefits and to minimize transaction costs. Yet, it in-evitably raises questions concerning the proper allocation of authority.National sovereignty has been eroded (Caporaso 1996; Hooghe 1996;Marks, Hooghe, and Blank 1996; Peterson 1997a; Risse-Kappen 1996;Sandholtz and Zysman 1989; Sbragia 1993; William Wallace 1996). Isthis outcome desirable, or should it be reversed? Should member statesand the Council of Ministers be strengthened, or should supranationalinstitutions such as the European Commission and the European Parlia-ment be bolstered? In the language of students of European integration,should the EU be intergovernmental or supranational ? This echoes debatesabout federalism in societies with center–periphery tensions (Elazar 1987;Sbragia 1992, 1993).

What should be the scope of authoritative regulation?

Should the European Union promote market liberal capitalism, based onthe Anglo-American model, or should it support regulated capitalism,rooted in the continental European Rhine model (Hix 1999; Kitschelt,Lange, Marks, and Stephens 1999; Rhodes and van Apeldoorn 1997;Streeck 1996, 1998; Streeck and Schmitter 1991; Wilks 1996)? The con-ventional left/right cleavage, which emerged out of the industrializationof Europe in the nineteenth and twentieth centuries, reflects contrast-ing views on the role of the state in the economy (Lipset and Rokkan1967; Page 1995). This cleavage, the most widely present fissure acrossEuropean societies, has now spilled over into the European arena.

What should be the principles of authoritative decision-making?

Should the European Union be democratic, like its constituent memberstates, or should it be technocratic, like other international organizationsfor economic cooperation? Political contention on this question goes backto the mid-1960s, when Jean Monnet’s method of piecemeal problemsolving was thwarted by French president Charles de Gaulle, partly on


grounds of defending national democracy against supranational techno-cracy (Dinan 1994; Duchene 1994; Hooghe and Marks 1999). Until themid-1980s, decision-making was elite dominated, but it has since beenprized open by wide-ranging public debate (Greven and Pauly 2000;Schmitter 1996, 2000; Helen Wallace 1993, 1996).

How should EU public interest be reconciled with national pressures?

The European Union is a heterogeneous polity. It knits together sometwenty nationalities; it mobilizes competition and cooperation among adiverse array of societal groups; and it creates interaction among social,economic, and territorial institutions from fifteen countries. How shouldCommission officials conceive of the public interest? Should they con-sider themselves architects and guardians of a common European interestthat transcends national, cultural, social, and territorial particularities?Or should the EU authorities act as agents of particular stakeholders inEuropean policies – public and private interest groups perhaps, and, lastbut not least, national interests? Should an elite speaking for the generalEuropean interest lead Europe, or should its elites be responsive to con-tending national interests? The debate between these conceptions echoesthe tension in national states between demands for pluralistic respon-siveness to societal interests and a more dirigiste state tradition, in whichbureaucracies are viewed as valuable instruments for continuity in a fluc-tuating political environment (Aberbach, Rockman, and Putnam 1981;Dogan and Pelassy 1984; Meny 1993; Page 1985, 1995; Suleiman 1975,1984).

These questions are fundamental for all democratic polities, and theyframe debate in the European Union. They bear directly on the roleof the Commission and its top officials. And in turn, top Commissionofficials, by virtue of their considerable powers in EU decision-making,are central voices in that debate. Yet we have had no clear idea about thepreferences of top officials.

Researching preferences

Preferences pose a serious research challenge to social scientists. Theycannot be observed objectively, unless one engages in in-depth structuredinterviewing. How have social scientists tackled this problem?

One response has been flatly to reject the significance of preferences. Inthe 1970s and 1980s, research on political preferences was largely aban-doned because they rarely seemed to predict behavior (Searing 1991).


As Robert Putnam once put it, “values and beliefs [were] discarded frompolitical analysis as froth on the mouth of madmen or froth on the wavesof history” (Putnam 1976: 103). The psychoanalytical school perceivedvalues and beliefs merely as rationalizations for emotional impulses, whileMachiavellians or Marxists claimed they simply cloaked self-interest orclass interest. However, the perceived gap between behavior and prefer-ences has less to do with perfidious or contradictory behavior on the partof the objects of study – human actors – than with poor conceptualizationand weak methodology on the part of social scientists. For example, manystudies focus on case-bound or time-specific preferences rather than onbasic preferences, that is, generalized beliefs and evaluations about socialand political life (George 1979). That is why this book does not focuson top officials’ preferences in relation to the Commission’s most recentanti-trust decision or the European Parliament’s vote on the reform of thestructural funds. Instead, it seeks to understand their basic preferenceson the four questions of European governance listed above.

Some scholars are skeptical about the value of researching preferencesbecause they harbor unrealistic expectations about their causal power.They place the bar too high. Preferences are general guidelines – heuristicaids to action – not a set of algorithms. They are context-sensitive propen-sities to action. Preferences should therefore be placed in a causal chainthat includes situational, institutional, and interactional factors (George1979; Putnam 1973; Scharpf 1997b; Searing 1994). Basic orientationsserve as “bounded rationalities” or “prisms” through which individualsconceive and respond to objective facts. They profoundly influence, butdo not determine, action.

This book describes and explains the basic preferences of top Com-mission officials. My point of departure is that this is an empirical ratherthan a deductive endeavor. If preferences help to shape action, then itmakes little sense to infer them from action.3 It is better to research pref-erences as directly as possible. Structured elite interviews enable one togain information about preferences that is independent of how actors be-have. I chose to ask top Commission officials – more than 200 of them –a common set of questions designed to capture basic dimensions of EUgovernance. Of these, 137 obliged.

This is an empirical enterprise, but it raises conceptual issues of gen-eral interest to comparative politics. What motivates elite actors to take

3 Some researchers have been tempted to shortcut the measurement problem by inferringpreferences from action. This is the approach implied by the notion of “revealed pref-erences” in economics. As Fritz Scharpf notes, “whatever may be its status in economictheory, if used as a methodological precept in empirical policy research, it could produceonly tautologies instead of explanations” (Scharpf 1997b: 60).


certain positions? How do they form preferences? Most social scienceexplains actions, and, in doing so, explanations often assume preferencesto be given – exogenous. But this begs the question why humans have cer-tain preferences in the first place. Understanding the sources of humanmotivation – the forces that shape human preferences – is the subject ofan intense debate that spans diverse theoretical approaches. My aim is tocontribute to a theory of preference formation.

Interests and values

EU governance has calculable consequences for the role of the Commis-sion, and for top Commission officials’ jobs. Top officials find it difficultto keep silent on the chief issues facing the Euro-polity because they arepoliticians; their power of initiative involves them intimately in politicaldecisions. Yet, as unelected officials appointed for their expertise, they arepresumed to be above the political fray. If they do not handle these ten-sions well, they may weaken their legitimacy. Commission officials rarelyadmit it, but their professional lives are at stake in the changing EU polity.It seems sensible for them to take these implications to heart when form-ing preferences on EU governance. This is what rational choice wouldpredict because it expects rational individuals to act and think in waysthat maximize their self-interest.

The idea of the rational, self-interested person has deep roots in Westernpolitical thought, and it is a key assumption underlying much social sci-ence (Mansbridge 1990). Yet we know that human motivation is complex(Elster 1989, 1994). There are many occasions when individuals’ pref-erences cannot be explained by self-interest – unless one is willing tostretch the concept of “rational self-interest” beyond common usage.4

Why do some rich people vote for leftist parties that favor income re-distribution? What explains the anonymous bravery of thousands who,during World War II, rescued Jews from Nazi persecution in Europe?5

4 On conceptual stretching, see Sartori (1991) and Collier and Mahon (1993).5 In her book on altruism, Kristen Renwick Monroe seeks to shed light on what motivates

ordinary people to risk their lives for others (Monroe 1996). One of her case studiesconcerns rescuers of Jews under Nazi occupation. She finds that altruists share a parti-cular “perspective, a feeling of being strongly linked to others through a shared humanityand [this] constitutes such a central core to altruists’ identity that it leaves them withno choice in their behavior when others are in great need” (Monroe 1996: 234). Theyadhere, in other words, to values that predispose them to adopt a universalistic worldview.Conventional rational actors do not hold these values. Monroe shows that altruismis not simply a function of socio-cultural or socio-demographic characteristics, or ofan economic incentive structure. Her study contributes to the debate about humanmotivation in that it demonstrates that self-interest is ill suited to explain some humanmotivations such as altruism.


More often still, rationality – understood as maximizing self-interest –fails to account for variation in the preferences of individuals with similarinterests. Why, for example, do Swedes, Danes, Norwegians, or Finnsgive five to seven times more of their income to developing countriesthan do equally wealthy US citizens?6 And why are farmers across theEuropean Union invariably among the most Euroskeptic, even thoughthey have been by far the greatest beneficiaries of EU largess over thepast fifty years (Gabel 1998c)?

There is more than one way to bring order to, or categorize, the richpalette of human motivations. Jon Elster, for example, distinguishes threesources of human motivation: rationality (or self-interest), emotions (orpassions), and social norms (Elster 1994: 21). In similar vein, DonaldKinder argues that the primary ingredients of political preferences arematerial interests, group sympathies and resentments, and principles(Kinder 1998: 800). People have preferences because they see benefit indoing so, because they feel emotionally compelled to, or because they be-lieve it is the proper thing to do. Other researchers propose more complexcategorizations (for discussions, see Chong 2000; Sears and Funk 1991).These categorizations can be grasped in terms of two basic contendingtheories of human motivation, an economic model that emphasizes utilitymaximization as mechanism for preference formation, and a sociologicalmodel that stresses socialization.

Utility maximization versus socialization

Utility maximization. Utility theory maintains that individuals aremotivated to maximize their utility and that they adjust their preferencesaccordingly.

The notion that human behavior is governed primarily by utility max-imization is particularly strong in economics, and it has gained groundin political science under the denomination of rational choice or publicchoice (Downs 1967; Moe 1990, 1997; Niskanen 1971; see also Chong2000). There is much debate among rational choice scholars about whatconstitutes maximization and what can be put into the utility function,

6 Scandinavian countries give on average 0.74 percent of their GNP as official aid to thedeveloping world – ranging from 0.33 percent in Finland to 1.01 percent in Denmark –against 0.11 percent for the United States. These figures for 1997/98 are drawn fromdata on aid to third world countries published every year by the Organisation for Eco-nomic Co-operation and Development (OECD). There is a widespread belief that theUnited States does badly in governmental aid but makes up for that through generousnon-governmental charities. But this perception does not have a basis in reality. TheOECD figures for non-governmental aid show that the United States is on a par withScandinavian countries: 0.03 percent of GNP against an average of 0.025 percent forthe Scandinavian countries.


but two principles seem relatively uncontroversial. The first is rationality,that is, individuals make decisions consistent with reasonable calculationsof costs and benefits. The capacity to calculate may be limited by theamount of information available and by external constraints, but the keyis that individuals act consistently in relation to their calculations. Thesecond is egoism, that is, that outcomes for oneself weigh more heav-ily than outcomes for others. However, some utility theorists stretch thedefinition beyond self-interest.

There is no consensus on a third idea, material hedonism, which saysthat utility maximization depends primarily on the acquisition of materialgoods (Sears and Funk 1991). This is a restricted notion of what individ-uals put in their utility function: it includes material interests, especiallymoney, but not non-material interests, such as social prestige or feelingsof moral contentment (Chong 2000; Elster 1989; Levi 1997b; Sears andFunk 1991).

When one conceives of human motivation primarily in terms of max-imizing utility, then this has clear implications for one’s understandingof preference formation. First of all, utility maximization presumes thatindividuals act consciously and deliberately. They calculate the costs andbenefits of alternatives, and any choice they make is intentional. In JonElster’s words, “rationality is a variety of intentionality” (Elster 1989: 7).Secondly, utility maximization assumes that rational individuals show noloyalty to existing norms, practices, or preferences when these no longerserve their interests. They adapt preferences to external constraints, andthey do so because they expect that such an adaptation will have desiredconsequences, that is, it will make it easier for them to realize their goals.In the most general way, actors want to make the best of a constrainedsituation. Unlike Don Quixote, they are reasonable in that they do notseek to tilt at windmills. This reflects what John March and Johan Olsenhave called “the logic of consequentiality” (March and Olsen 1989: 160).Thirdly, it follows from this that political preferences are not very stable.Individuals may update their preferences quite frequently, depending onthe incentive structures they face.

Socialization. The core idea of socialization theory is that indi-viduals acquire preferences by internalizing norms, values, and princi-ples embodied by the groups or institutions that are important to theirlives. This view emphasizes the centrality of group ties and longstandingpersonal dispositions, which organize individual preferences and shapebehavior (Converse 1964; Johnston 1998; Kinder 1998; Rohrschneider1994, 1996; Searing 1969, 1986; Searing, Wright, and Rabinowitz 1976;Verba 1965).


The notion that group ties and values – longstanding, deeply ancho-red dispositions – govern human behavior has a strong foundation insociology as well as in psychology (Glenn 1980; McGuire 1993). It is thedominant paradigm among students of political attitudes and preferences(Kinder 1998). How these dispositions become anchored as part of anindividual’s identity and how they shape this person’s preferences and be-havior towards new objects or issues are matters of debate. Yet scholars inthis socialization paradigm concur in seeking explanations in sociologicaland psychological processes – not in human instinct for rationality.

Individuals develop preferences for social reasons that cannot be at-tributed to direct material incentives or coercion. The absence of a strongassociation between self-interest and preferences is the key differencebetween a socialization and a utility maximization model. Socializationdirects the individual away from the self to the socializing group. Theoriginal meaning of the word “to socialize,” after all, is “to render social,to make fit for living in society [or in the particular group]” (Conover1991: 131).7 This notion of a non-material, social basis of preferences iscrucial for the socialization perspective.

Another widely shared notion is that socialization is a gradual, in-cremental process. Individuals typically internalize values or preferencesthrough innumerable encounters with particular political norms or prac-tices. Socialization takes time (Verba 1965). The longer one is exposedto particular stimuli, the more one is likely to absorb these influences(Kinder and Sears 1985; Glenn 1980).

To be sure, socialization is not always gradual. It is occasionally broughtabout by dramatic single episodes. One-time, powerful events – wars,historic elections, or protests – may jolt individuals into changing political

7 A critical issue in the socialization literature, which bears on my topic indirectly, concernsthe micro-processes of socialization. Through what processes are norms and values trans-mitted? Socialization scholars generally distinguish four processes, which range from theself-conscious to the subconscious. At one end of the continuum stands persuasion, orsocial learning, whereby individuals are convinced through self-conscious cognition thatparticular norms and causal understandings are correct and ought to guide their ownbehavior. Second, social influence refers to the process whereby individuals’ desire tomaintain or increase social status or prestige induces them to conform to group norms.The group rewards an individual’s behavior with back-patting and status markers or pun-ishes it with opprobrium and status devaluation (Checkel 1998; Johnston 1998). Somescholars are reluctant to consider social influence or pressure as a genuine socializationprocess because it does not necessarily require that individuals actually change their pref-erences. Social influence is “public conformity without private acceptance” (Johnston1998). A third process is social mimicking, whereby an individual “inherits” or copiesnorms and behavior without putting much conscious thought into it (Beck and Jennings1991; Johnston 1998). Finally, attitude crystallization refers to a largely subconsciousprocess whereby individuals extend deeply held, stable attitudes to new attitude objectsthrough cognitive or affective processes (Converse 1964; Sears and Valentino 1997).


preferences. The open, still unshaped minds of young adults appear to beparticularly susceptible to this type of socialization (Sears and Valentino1997). It is not impossible for adults too to be “shocked” by such events,and to embrace a different value system as a result of it. However, thisusually demands an “exacting and unusually powerful social situation”(Kinder and Sears 1985: 724).

There is less agreement on how internalized dispositions guide pref-erence formation on new political objects, that is, the psychological pro-cesses underlying political judgment. Some scholars conceive of theseprocesses as primarily cognitive, others as emotive/affective. The formerconceive of people as “cognitive misers” who have limited capacity forprocessing information and who therefore must use cues or prior knowl-edge to reach judgments on new objects (Conover and Feldman 1984;Taylor 1981). The latter scholars emphasize the emotive/affective aspectsof political judgment (Sears 1993; Sears and Funk 1991). Yet commonto these explanations is the psychological notion of belief or affect consis-tency, which maintains that individuals’ quest for consistency in beliefsor affects is a basic human desire (Sears 1993; Sears and Funk 1991).

The leading cognitive model, the on-line model, assumes that peo-ple keep summary evaluations (“running tallies”) of important politicalprocesses or principles. They use these to spontaneously evaluate newobjects and, although they regularly update their tallies, they quickly for-get the details that prompted the updating in the first place (Lodge andSteenbergen 1995). So dispositions – running tallies – are not immutablyfixed; they are incrementally updated, and yet they provide cognitive guid-ance to individuals in evaluating new experiences.

David Sears’ symbolic politics theory is the most prominent exampleon the emotive side (Sears 1993; Sears and Funk 1991). The modelholds that people acquire stable affective responses to particular symbolsthrough socialization.8 These dispositions influence preferences on newpolitical objects in the following way. Each political object is composed ofone or more symbols, and these symbols determine which dispositions areinvoked. Individuals automatically transfer the internalized affects fromthese invoked symbols to the new object.9

8 The strongest dispositions, namely party identification, political ideology, and racialprejudice, are called symbolic dispositions, and they may last a lifetime.

9 Cognitive and affective models also see eye to eye in that they emphasize the preva-lence of non-rational over rational processes. David Sears summarizes the contrast witha rational perspective: “The rational choice view depicts a deliberate decision makerobjectively evaluating costs and benefits. The cognitive miser approach sees a cerebralbeing desperately trying to husband his or her limited psychic energy in the midst ofa torrent of information. The symbolic processor is reacting in a gut-level, automaticmanner to emotionally evocative political and social objects” (Sears 1993: 137).


When one conceives of human motivation primarily in terms of livingby predispositions, this has clear consequences for preference formation.First of all, a socialization perspective presumes that individuals react in-stinctively to cognitive cues or affects. People follow established traditionsand conform to social norms in a relatively reflexive fashion. They do notcalculate probable costs and benefits (Sears 1993). Second, the centralityof socialization means that individuals show considerable loyalty to in-ternalized norms, practices, or preferences. Individuals hold on to these,even when they cease to serve their interests. What motivates them is, inMarch and Olsen’s terms, “a logic of appropriateness” (March and Olsen1989). They ask: Who am I? What group am I part of ? So what is theproper thing to do? They do not ponder on what is in it for them. Thatis why the controversy between the two paradigms is often presented asa conflict between values and interests (Chong 2000; Sears and Funk1991). The implication of this is that basic preferences are perceived tobe relatively stable, even rigid. They do not adjust easily to changing cir-cumstances; rather, individuals have a tendency to integrate new politicalobjects or situations into familiar value systems.

Utility maximization and socialization perspectives tap distinct, yet com-plementary sources of human motivations. Explanations for preferenceformation that privilege one above the other produce parsimoniousmodels, but the truth is that both interests (utility) and values (social-ization) are part and parcel of human life. Why would we, then, want toexclude one or the other?

A model of interests and values

For many years, the attempt to combine utility maximization and social-ization appeared to be as fruitless as combining water and fire. Theirontological and epistemological foundations seemed just too different.Yet, recent work by Dennis Chong has carved out a path for a genuinelyintegrated approach to preference formation (Chong 1996, 2000).

Chong, one of a handful of political scientists with active researchin both traditions, identifies some key weaknesses of each approach.Chong’s main critique of the socialization model is that it does not pro-vide a convincing account of value or norm formation, that is, how groupnorms originate and why they differ, why people conform to group norms,why groups put so much effort into imposing their norms on others,and when values or norms change (see Chong 2000: 36–7; 1996: 2082).Socialization theory emphasizes the generality and stability of attitudes,and underemphasizes conflict over ways of life. “A static model built on


relatively fixed dispositions encounters difficulty explaining changes innorms and values” (Chong 2000: 45). Chong makes a case for treatingnorms and values strategically, and he identifies ways in which inter-ests and rational decision-making support values and norms. His mainargument is that current dispositions are often based on past strategiccalculations. Or, to put it differently, the reasons for internalizing normsin the first place often have to do with rational interest.

In fairness to the socialization perspective, considerable work has beendone to specify the conditions under which internalized dispositions maychange (e.g. Glenn 1980; Searing, Wright, and Rabinowitz 1976; Searsand Valentino 1997). Research specifying the micro-foundations of thesocialization model, such as the on-line model and the symbolic politicsmodel, brings to light the psychological processes by which change indispositions occurs. However, these explanations rarely reserve a spacefor interest. Chong’s argument, in contrast, is that rational calculationenters into and shapes socialization.

Chong criticizes the rationality assumption in the utility model be-cause it does not specify how norms, practices, and values – internalizedthrough socialization – facilitate rational decision-making. The idea ofbounded rationality, popularized by Herbert Simon, creates an openingto bring values into rational decision-making. Bounded or proceduralrationality is relevant for “a person who is limited in computational ca-pacity, and who searches very selectively through large realms of possi-bilities in order to discover what alternatives of action are available. . .The search is incomplete, often inadequate, based on uncertain informa-tion and partial ignorance, and usually terminated with the discovery ofsatisfactory, not optimal, courses of action” (Simon 1985: 295). Humanprocessing of information under these imperfect conditions has been amajor topic in cognitive psychology (Kinder 1998 presents an overview).Chong, like cognitive psychologists, starts from the assumption that indi-viduals are fallible humans with myopia or plainly limited brains. In thiscontext, norms, practices, values, and group ties become indispensable.They provide powerful cues to help people figure out their utility andchoose among alternatives.

It is true that many rational choice analyses now routinely examinelimitations to available information and constraints on information pro-cessing. Yet norms, values, and dispositions appear in these models asdisturbing factors that derail rational calculation, or they are relegatedto a residual category (for discussions see Kato 1996; Nørgaard 1996;Weingast 1995; Yee 1997). Chong presents an argument in which normsand values become an integral part of rational calculation.

Chong goes on to develop a general model of preference formation“that combines a social psychological model in which identities and


values are socialized through socialization with an economic model inwhich belief and value formation are motivated by external or instru-mental benefits” (2000: 7). Preference formation depends on two fac-tors: (a) the costs and benefits (material and social incentives) of presentalternatives, and (b) dispositions, which have themselves been formed bypast investments in values, group identities, and knowledge. So short-term and long-term factors combine to explain a political preference ona particular object.

This model of preference formation constitutes a breakthrough. Itidentifies dependencies among apparently incompatible paradigms. Themodel also has a simple structure. Preference is the function of two in-dependent variables: strength and direction of dispositions for each al-ternative, and cost/benefit calculations for each alternative. The key task,then, is to estimate parameters for each variable. Furthermore, the modeloffers a plausible interpretation of complex human motivation. The so-ciological perspective conceives of individuals as social beings – hominessociologici – who live their lives to a large extent in the shadow of effectivesocial norms and values. Utility maximization suggests that individualsare self-interested beings – homines economici – who consciously calculatethe costs and benefits of their actions and preferences. In Chong’s model,the question is no longer whether individuals live in utilitarian cocoons orsociological cages (Searing 1991). Rather, the task is to understand theinterplay between socialization and utility maximization and to specifythe conditions under which the balance between the two varies. When,for example, do career concerns weigh more heavily than ideology fortop Commission officials on basic issues of EU governance? Are officialsmore likely to be mindful of their career when they choose between asupranational or intergovernmental Union than when the choice is be-tween a neoliberal or regulated European economy? Does it depend onwhere they work in the Commission, their nationality, and their length ofservice in the Commission? What makes values trump interests, or viceversa? These concrete questions raise conceptual and methodological is-sues, which now need to be addressed.

An empirical approach

How can one develop a scientific, that is to say falsifiable, theory of pref-erence formation?

Workable concepts. A testable, falsifiable model requires defini-tions that unambiguously separate key concepts from each other. Thisnecessitates some hard choices and, though these choices are inherently


debatable, they should also be transparent. A model that examines therelative impact of utility and socialization should avoid slippery concepts.

The trickiest issue concerns the boundaries of utility. What could arational individual reasonably seek to maximize? At the heart of utilitytheory stricto sensu is a definition of rationality that emphasizes the maxi-mization of economic self-interest, often expressed as money (Kato 1996;Yee 1997). In rational choice jargon, this is the thick version of rational-ity. It transfers the notion of rationality used in the neoclassical economicmodel to the analysis of politics. Rational interest is defined in terms of(a) the material wellbeing and (b) individuals’ own personal life (or thatof their immediate family). Both definitional conditions are contested.

Margaret Levi represents the opposite pole of the debate when sherecommends using a thin variant of rationality in which individuals actconsistently in relation to their preferences. In this view, any kind ofobject or value could be maximized, including economic self-interest,economic group interest, non-economic interests, power, security, statusand respect, ideas or norms, an interest in pleasure, or an interest indeveloping a coherent understanding of society. There is no reason, then,to stick to the assumption that individuals are self-interested; they couldseek to maximize the utility of their family, their community, their party,or their country. The only methodological requirement is that researchersshould define the value ex ante (Levi 1997b).

If one were to adopt a thick notion of utility, one would focus onwhether Commission officials could draw personal financial benefit froma particular attitude. Were one to adopt a thin notion of rationality, theutility function of Commission officials could encompass personal power,the power of the Commission, the status of their nationality, or maximiz-ing the European Public Good.10

A thin version of utility maximization can easily produce tautology:whatever people do becomes a revealed preference.11 A thin notion is alsoat odds with the common sense understanding of rational interest. More-over, by refusing to exclude the most unambiguous non-self-interestedgoods, it opens the door for nonsensical arguments in which heroic actsare explained in terms of altruism, or philanthropy is presumed to bemotivated by the desire to maximize the pleasure of being good. Thisconfounds values and interests.

10 For example, the two founders of public choice, Gordon Tullock (1965) and AnthonyDowns (1967), interpreted self-interest broadly to include suprapersonal values of vari-ous sorts, for example regarding good policy or the public interest (Moe 1997: 457 fn.1).As I argue below, this amounts to rather creative conceptual stretching of the neoclassicaldefinition of self-interest.

11 This is why Levi advises researchers to specify the utility function ex ante (Levi 1997b:24).


It seems therefore sensible to limit utility maximization to materialinterests. Many rational choice analysts restrict utility to an individual’sown material interest. However, I propose to broaden this to includematerial group interest. It seems reasonable to extend the motivation ofutility maximization to individuals acting on behalf of their groups whenthese individuals share in expected group benefits.

Where does this leave us? Following Levi’s suggestion, I define ex antewhat is included in the utility function of top officials. The single mostimportant thing that should jolt a top official into strategic preferenceformation is a concern for his (or her) professional career. Top officialswork in an environment where competition for attention from principals,agenda setting, resources, prestige, influence, and promotion is harsh. Tobe ignored or bypassed by their principals once or twice may publicly taintan ambitious official; to be demoted to a peripheral service is equivalentto semi-permanent exile; to be sidetracked to an advisory position offthe normal hierarchical line usually means premature career death. TopCommission officials cannot easily be fired, and they earn very handsomesalaries. Yet, barring these, there is considerable scope for the College ofCommissioners to affect top officials’ professional careers and so onemay expect rational officials to take this into account when forming theirpreferences.

My definition of utility maximization focuses on material interests be-yond the neoclassical concern with individual wealth maximization. Iexpect rational Commission officials to be motivated primarily by indi-vidual career concerns. Yet I think it also likely that they may maximizegroup benefits. For example, it seems sensible to categorize a top officialwhose preferences reflect opportunities to maximize economic payoffsfor his country as motivated by utility maximization. After all, he canreasonably expect that he (and his family) will share in the benefit.12 I donot wish to rule out a link between material group interest and materialindividual interest.

Defining utility maximization broadly has a strategic advantage for hy-pothesis testing. It gives utility factors the greatest possible chance to bepicked up in a contest with socialization variables. This is important be-cause socialization theory has long been the dominant paradigm in thestudy of political attitudes and preferences (Chong 2000; Kinder 1998).Utility maximization is a newcomer.

12 The reader may be surprised that I mostly use masculine pronouns when referring toCommission officials. To use “she or he” would create a false impression of genderbalance in an institution where only 6.5 percent of top Commission officials are female.Instead, I employ feminine pronouns only occasionally to reflect the extreme genderimbalance in the top layers of the Commission bureaucracy.


Operationalizing socialization is more straightforward. Socialization the-ory suggests that a top official’s preferences on EU governance reflectinternalized norms, values, and principles embodied by the groups orinstitutions that have been important in his life.

Most theorists conceptualize socialization as a gradual process thattakes considerable time. They also expect experiences in childhood oryoung adulthood to be more formative than experiences later in life, forexample at mid-career. Yet it is possible that “exacting and unusuallypowerful” (Kinder and Sears 1985: 724) one-time events may jolt topofficials into internalizing particular preferences, although these occur-rences should be rare. For example, one might expect the resignationof the Commission in 1999 to constitute such an event. However, so-cialization requires, most of the time, sustained exposure to consistentlytransmitted norms and values. Socialization variables should therefore besensitive to time.

Utility maximization presumes that top officials are guided by tangible,immediate career benefits, whereas socialization implies that top officials’preferences reflect intangible, longstanding dispositions that were inter-nalized through participation in groups or institutions over time. Thisway, the two basic logics are unambiguously distinct.

Testable hypotheses. It is one thing to be able to recognize alter-native human motivations when one encounters them. The next step isto hypothesize circumstances in which one may expect socialization orutility maximization to shape top officials’ preferences on basic issues ofEU governance.

My guiding hypothesis is that top officials’ preferences are rooted in ex-perience. However, not every experience shapes their calculations or dis-positions to EU governance. What is needed is an approach that enablesone to distinguish relevant from irrelevant experiences. Of the many op-portunity structures encountered by top officials, which induce them toadjust their preferences to further their career? And which internalizedvalues are likely to shape their preferences on EU governance?

A focus on institutional rules allows one to think systematically aboutthe contexts in which utility or socialization take place.13 The conceptof institution or institutional rules is simpler, yet more general, than

13 This ties in with the resurgent study of institutions in rational choice and the literature onsocialization and learning. Rational choice institutionalism and historical or sociologicalinstitutionalism conceive of individuals as rule bound (for overviews, see Aspinwall andSchneider 2000; Hall and Taylor 1996; Ostrom 1986, 1991; Thelen 1999). That is,in explaining preferences and behavior, they give analytical priority to the particularinstitutional rules that constrain individuals at a given time.


opportunity structures, incentive structures, reference groups, socializa-tion agents or schools, churches, and family – often employed by utilitymaximization or socialization perspectives to describe particular contexts.

What are the relevant institutional contexts for top Commission offi-cials? It seems reasonable to expect top Commission officials to be influ-enced by multiple institutional contexts simultaneously. Why is that so?

In a simple world, a single institution would mold top officials’ pref-erences. If the Commission were a strongly bounded institution withperfect control over top officials’ working life, it would be Commissioninterests and values that shape its employees’ preferences. The questionwould then be how top officials acquire Commission preferences. Is itbecause the Commission is an effective socialization agent (a process ofvalue transfer), or because it controls top officials’ careers (a process ofstrategic calculation and adjustment)?

Yet the Commission lacks the insulation to mold its employees’ pref-erences uniformly. Top Commission officials have diverse cultural andeducational backgrounds and very different professional experiences. Sothey start as a diverse bunch; and the Commission’s way of workingpermits them to remain heterogeneous. The Commission is a compart-mentalized bureaucracy, where many directorates-general resemble self-governing statelets. This makes it possible for top officials – the bosses ofthese statelets – to mold the norms and habits of their own small world totheir own image, and thus to persevere in being different. Socializationscholars tell us that such institutional conditions are not conducive tocreating a homogeneous, single-purposive service. This contrasts sharplywith many national administrations in Europe.

Second, the Commission’s grip on the professional lives of its top em-ployees is subject to national governmental and, in a number of cases,partisan control. National governments or political parties can and doinfluence Commission officials’ careers indirectly. They monitor closelywhether the Commission abides by the agreed national quota for seniorpositions. They often draw up a shortlist of candidates for high-level po-sitions that are reserved for candidates from outside the Commission(parachutage), and they may have de facto “right of first refusal” for suchpositions. Finally, they usually enjoy ties with the personal cabinets of thecommissioner of their nationality or party allegiance, and these cabinetsare the brokers in Commission personnel policy. These flaws in Commis-sion control over its own employees do not induce rational top officialsto adjust their preferences to the Commission’s, as a cursory analysis oftheir utility function would suggest.

Third, the wider political context influences top Commission officialsdeeply. As noted above, Commission officials are players in a multi-level


political system in which authority is shared vertically across territoriallevels and horizontally between several EU and national institutions. Inmany (now most) policy areas, no institution – not the Commission orthe Council or the European Parliament, not national governments –can take authoritative decisions unilaterally (Hooghe and Marks 2001,appendix 1; Pollack 1995, 2000; Schmitter 1996). Top officials need tobe attuned to national governments, political parties, and the public, aswell as to the European institutions – and not only, or even primarily, tothe Commission. The European system of multi-level governance plugstop Commission officials into multiple institutional contexts.14

These observations shatter the notion of the Commission as a unitaryinstitution capable of controlling the preferences of its employees. Whatcomes into view is a multi-layered institutional context, where institutionssmaller than the Commission, such as the cabinet or the directorate-general (DG), loom larger, and institutions outside the Commission,such as political parties, national administrations, or national political sys-tems, come closer. Top Commission officials’ preferences on Europeangovernance are likely to be shaped by experiences inside and outside theCommission.

We now have the building blocks for a testable model of preference for-mation in the Commission. By combining type of institutional contextwith logic of influence, it is possible to pin down a limited number ofplausible influences on top Commission officials’ preferences. These arepresented in figure 1.1.

Preference formation in the Commission

Figure 1.1 summarizes how the life and career paths of top officials mayshape their preferences on EU governance. The cells in figure 1.1 applysocialization and utility maximization to institutional contexts inside theCommission and institutional contexts outside the Commission. I havetwo general expectations:

14 There is a growing literature on how a system in which authority is diffused across ter-ritorial levels – a multi-level polity – influences the preferences and behavior of publicopinion (Anderson and Gabel 2000), political parties (Bomberg 1998; Hix 1999; Marksand Wilson 1998, 2000), trade unions (Ebbinghaus 1999; Ebbinghaus and Visser 1997;Turner 1996), social movements (Imig and Tarrow 1997, 2001; Marks and McAdam1995; Tarrow 1995, 1999), firms and business representation (Coen 1997; Greenwood1997), national and regional governments (Hooghe 1996; Marks 1996a,b; Marks et al.1996), structures of interest intermediation (Falkner 1996, 1999), and policy networks(Kohler-Koch and Eising 1999). It seems reasonable to assume that top Commissionofficials’ desires and deeds too should be affected by their involvement in various insti-tutional settings of this multi-level system.


Figure 1.1 Top officials and preference formation on EU governance.

1. Basic preferences on EU governance are likely to reflect a mix of so-cialization and utility factors. Most people are, most of the time, bothrational and moral.

2. These preferences are likely to be shaped by experiences outside aswell as inside the Commission. The Commission and its components(DG and cabinet) constitute merely one part of a multi-faceted in-stitutional environment, which also includes national political parties,national administrations, national networks for socializing in Brussels,or national political systems. Let us examine these.

Commission – socialization vs. utility maximization

We have unmasked the Commission as a weak institution, but it wouldbe rash to dismiss it altogether as a potential influence on top officials’preferences. A typical top official clocks workdays of close to ten hours,five or six days a week, and that should give plenty of time for socializationor utility calculation.

The prevailing assumption in much work on European integration isthat the Commission is pro-integrationist. This assumption character-izes the first major theory of European integration, neofunctionalism.

Documents

TheEuropeanCommissionand the Integration of Europe